|Publication number||US7506315 B1|
|Application number||US 10/750,044|
|Publication date||Mar 17, 2009|
|Filing date||Dec 30, 2003|
|Priority date||Dec 30, 2003|
|Publication number||10750044, 750044, US 7506315 B1, US 7506315B1, US-B1-7506315, US7506315 B1, US7506315B1|
|Inventors||Mario Kabadiyski, Nikolai G. Nikolov|
|Original Assignee||Sap Ag|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (19), Non-Patent Citations (38), Referenced by (51), Classifications (10), Legal Events (3)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
The field of invention relates generally to testing and monitoring of program code; and, more specifically, to an application tracing service which employs bytecode modification techniques for testing and monitoring of program code.
2. Description of the Related Art
Certain software technologies, such a Java, emphasize the use of a special interpreter, referred to as a “virtual machine,” that allows generic processor instructions to be executed on a particular type of processor. Here, each hardware platform (e.g., each computer) that the generic instructions are expected to “run on” typically includes a virtual machine interpreter that is responsible for converting the generic processor instructions, referred to generally as “interpreted code” or “bytecode” in the case of Java, into code that is specially targeted for the hardware platform's particular processor. Software technologies that embrace the execution of bytecode on a virtual machine may be referred to as “virtual machine-based” software.
As a classic instance of the benefit of the Java virtual machine with respect to Internet usage, a first PC that is powered by an Intel processor may download from the Internet the same Java bytecode instructions as a second PC that is powered by a PowerPC processor. Here, the first PC's Java virtual machine converts the Java bytecode into instructions that are specific to an Intel processor while the second PC's Java virtual machine converts the same Java bytecode into instructions that are specific to a PowerPC processor. Thus, through the use of Java bytecode and processor specific Java virtual machines, an Internet server is able to maintain only a single type of code (the Java bytecode) without concern of client compatibility.
2. Object Oriented Programming and Classfiles
Certain software technologies, including Java, are “object oriented.” According to an object oriented approach, the subject matter that is processed by a computer program is organized into classes of likeness. For example, the software used to sell items to customer X might belong to the same class of software (i.e., the class: “sales”) that is used to sell items to customer Y. Here, given that a significant degree of overlap is expected to exist regarding the methods and data types used to process sales for both customers X and Y (e.g., Part Nuumber (PN), Quantity (QTY) “update billing about sale,” “update accounting about sale,” . . . etc) it is deemed more efficient to organize such methods and data types into a generic “sales” class from which specific instances of the class (e.g., an instance for selling to customer X and an instance for selling to customer Y) can be defined and created.
Each specific instance of a class is referred to as an object; and, each object assumes the characteristics of the class from which it is defined. Thus, a first object could be created from the generic sales class to create a “sell to customer X” object; and, a second object could be created from the generic sales class to create a “sell to customer Y” object. Both objects would be able to execute the methods defined by the class. For example, the “sell to customer X” object could execute the “update accounting about sale” method to update an accounting department about a sale to customer X; and, the “sell to customer Y” object could also execute the “update accounting about sale” method to update an accounting department about a sale to customer Y.
In general, the runtime execution of any computer program can be viewed as the execution of a sequence of methods. With respect to an object-oriented approach, such a sequence of methods can be implemented by calling upon a sequence of objects and invoking one or more methods at each object. In order to invoke the method of an object, a representation of the object must first be created. In virtual machine-based object-oriented software environments, classfiles are bytecode level data structures from which such representations are created. A classfile can be viewed as a bytecode level description of a class; and, therefore, a classfile can be used as a template for the formation of an object, at the bytecode level, that is a member of the class. As specific methods can be attributed to specific classes, each classfile is made to contain the sequence of bytecode instructions that correspond to each of the class's specific methods.
Compilation 102 of the source code level class description 101 b results in the formation of the bytecode level classfile 103. Among other possible items not shown in
Over the course of discussion of various inventive aspects set forth in the detailed description that follows, comparisons will be made against each of
Note that both the source code level and bytecode level implementations for the “GetMax” method have a single entry point 140 a, 140 b (i.e., the method starts at locations 140 a, 140 b) and a pair of exit points 141 a, 142 a and 141 b, 142 b (i.e., the method can end at locations 141 a, 142 a and 141 b, 142 b—noting that a “return” command/instruction causes an output to be presented; which, in turn, can be viewed as the completion of the method). Those of ordinary skill will be able to recognize that: (1) the source code level depiction 120 a of the GetMax method observed in
A method having a single entry point 140 and a pair of exit points 141, 142, like the “GetMax” method, can be represented by depiction 150 where the top surface 140 corresponds to an entry point of the method and side edges 141 and 142 correspond to exit points.
Entry points are depicted at the top surfaces 209, 210, 211, 212 of methods 205, 206, 207, 208; and, exit points are depicted as side ports of methods 205, 206, 207, 208 (i.e., side ports 213, 214, 215 for method_1 205; side ports 217, 218 for method_2 206; side ports 220, 225 for method_3 207; and, side ports 221, 222, 223, 224 for method_4 208) and bottom ports of methods 205, 206 and 208 (i.e., bottom port 216 for method_1 205; bottom port 219 for method_2 206; and, bottom port 226 for method_4 208).
According to the object oriented process flow depiction of
3. Enterprise Systems
Traditional client-server systems employ a two-tiered architecture such as that illustrated in
As is known in the art, the “business logic” component of the application represents the core of the application, i.e., the rules governing the underlying business process (or other functionality) provided by the application. The “presentation logic” describes the specific manner in which the results of the business logic are formatted for display on the user interface. The “database” 232 includes data access logic used by the business logic to store and retrieve persistent data.
The limitations of the two-tiered architecture illustrated in
In response to limitations associated with the two-tiered client-server architecture, a multi-tiered architecture has been developed, as illustrated in
This separation of logic components and the user interface provides a more flexible and scalable architecture compared to that provided by the two-tier model. For example, the separation ensures that all clients 230 share a single implementation of business logic 244. If business rules change, changing the current implementation of business logic 244 to a new version may not require updating any client-side program code. In addition, presentation logic 242 may be provided which generates code for a variety of different user interfaces 240, which may be standard browsers such as Internet Explorer® or Netscape Navigator®.
The multi-tiered architecture illustrated in
Although the multi-tiered system illustrated in
A system and method for coupling a distributed statistical records (“DSR”) collection system with a tracing system utilizing bytecode modification techniques. In one embodiment, a distributed statistical records (“DSR”) module initially identifies one or more application components within the application server to be traced. A bytecode modification module responsively modifies the bytecode of the one or more application components. The modifications are directed to a particular set of methods of the application components (e.g., methods which are relevant to DSR data). A dispatch unit registers the method invocations and method-related information associated with the particular set of methods and provides the method-related information to the DSR module. Finally, a DSR interface module translates the method-related information to a format employed within a DSR system and forwards the translated information to the DSR system.
A better understanding of the present invention can be obtained from the following detailed description in conjunction with the following drawings, in which:
1.0. Architectural Overview
Developers, based on their particular testing, debugging and/or monitoring needs, may customize design and implement plug-in modules. This modular plugin architecture provides a flexible and scalable solution for evaluating program code, and may be particularly useful for evaluating program code on large enterprise platforms (e.g., such as J2EE-based platforms).
One particular type of function performed by a plug-in handler is an output function that records one or more properties about the method and/or causes one or more properties associated with the method to be recorded. For example, one type of output function performed by a plug-in handler (e.g., plug-in handler 331 1) involves recording the time that each method started at an entry point and/or exited at an exit point. Developers may use this recorded timing information to determine, for example, the sequence of methods just prior to a “crash” of the software. Similarly, if a particular method or sequence of methods take an inordinate amount of time to complete, the developer may determine the reasons why and/or change the program code to operate more efficiently. As such, the recording of when method entry points and method exit points are reached allows a precise method history sequence to be reconstructed.
In addition to timing information, the parameters that are passed to a method when it starts (e.g., its input parameters) and/or the parameters that are passed by a method when it exits (e.g., its output/returned values) may also be recorded by a plug-in handler to enhance the recorded history of the execution flow (e.g., a returned value or a thrown exception). For example, applications may often crash because of an incorrect or incompatible parameter value. By recording parameter values that are passed in and out of the application's methods over the course of the application's execution flow, the appearance and/or source of an incorrect/incompatible parameter value can be readily flagged.
Thus, not only can start/stop times be recorded on a method-by-method basis; but also, parameters can be recorded on a method-by-method basis. Of course, the identity of the methods themselves may also be recorded. A plug-in handler that records the start and/or end times of a method and/or records the parameter flow in and/or out of a method may be referred to as a “tracing” plug-in handler because aspects of the method history of a software routine can be “traced” through the recordation of its method start/end times and its method parameter values.
Another type of output function that may be performed by a plug-in handler (e.g., plug-in handler 331 2) involves maintaining a counter for each bytecode modified method and incrementing the counter each time its corresponding method is used (e.g., incrementing the counter for a bytecode modified method each time the bytecode modified method makes a function call to the dispatch unit 330 from its entry point). As applications (and, in particular, enterprise applications) involve complicated “logic,” cross-method interdependencies, etc., it is possible that certain sections of code are never executed (i.e., are “dead” or “unreachable”) or are almost never executed because of a flaw in the design of the application.
Counting each time a method is used allows dead/unreachable or hardly used regions of the software to be readily identified. In addition, for monitoring purposes, counting each time a method is used allows a software manager to comprehend “usage” rates of the software's individual methods. A plug-in handler that is capable of maintaining a counter for an individual method may be referred to as a “coverage” plug-in handler because the frequency of usage of various regions of the software can be gauged by counting each time a method is used. In related embodiments, a single coverage plug-in handler may be designed to maintain separate counters for a plurality of methods.
In the examples described herein, the various functional components illustrated in
2.0. Modified Bytecode Runtime Flow
Once the methods to be modified have been modified 452, the modified bytecode is executed 453 with the help of a virtual machine interpreter. The execution of the bytecode may also be referred to as “runtime”. During runtime 453, as explained in more detail further below, the dispatch unit 330, 430 responds to each invocation from a modified method by ensuring that the runtime execution process flows to the appropriate plug-in(s) for the invoking method. As a consequence, the invocations made to the dispatch unit 330, 430 cause the appropriate plug-in handler(s) to perform their designated functions (e.g., such as the output functions described above). Note that each plug-in handler can be viewed as including a specialized body of instructions that, when executed, cause the plug-in handler's particular function to be performed.
For the purpose of illustration, one example of system operation is set forth in
Thus, as a consequence of the bytecode modification, for a process flow that performs the substantive equivalent of that observed in
Note that as bytecode modifications are made at each entry and exit point for each of methods 405-408, the associated plug-in handlers may still record information for each of methods 405-408 even if the runtime process flow were to flow through a different set and/or sequence of method entry and exit points than those observed in
In the particular example illustrated in
By contrast, the process flow for method 407 involves from its entry 411 and exit 420 points: (1) invoking dispatch unit 430; (2) dispatching the invocation to plug-in module B 470; (3) executing tracing method(s) with handler 462 (i.e., record a start time and/or input parameters if invocation is from entry point 411; or, record an end time and/or output parameters if invocation is from exit point 420); (5) executing coverage method(s) with handler 463 (i.e., increment a counter maintained for method_3 407) and, (5) return back to the entry or exit point that made the original invocation.
The above example demonstrates that the embodiment of the dispatch unit 430 referred to above is responsible for being conscious of an appropriate plug-in module for each invoking method and causing the runtime process flow to be directed to the appropriate plug-in module for each invoking method. As observed in
Note that appropriate coverage monitoring may be gained by the coverage plug-in handler 463 at either a method's entry point or a method's exit point (e.g., rather than executing the coverage plug-in handler 463 at both the method's entry and exit points as suggested by
In response to the invocation, the dispatch unit 430 identifies the appropriate plug-in module for the method through recognition of the invoking method (e.g., by way of a class ID and a method ID) and causes the process flow to be dispatched to an appropriate plug-in module 503 by returning a reference to the appropriate plug-in module for the invoking method (e.g., a reference to an object that the plug-in module corresponds to). The invoking method uses the reference to call upon the plug-in module; which, in turn, causes the plug-in module's handler method(s) to be executed. The handler(s) execute their recordation method(s) 504; and, the process flow returns to the remainder of the invoking method's instructions 505. In the case of
According to the methodology of
In response to the invocation at the method's exit, the dispatch unit identifies the appropriate plug-in module for the method through recognition of the invoking method and provides back to the invoking method a reference to the appropriate plug-in module. The invoking method uses the reference to call the appropriate plug-in module 508. The plug-in module's handler(s) execute their recordation method(s) 509; and, the process flow returns to exit the method.
3.0. Example of Modified Bytecode
For additional information and ease of understanding, an equivalent source code method 620 a that could be compiled to directly produce the bytecode modified 620 b version is also shown in
Comparing the pre-modification bytecode method 120 b of
Blocks of instructions 643 b through 646 b are shown in
This may involve, for example, the placing of these arguments on top of an operand stack. In a further embodiment, because the dispatch unit 430 may also be geared to recognize the invoking method based upon (at least in part) the invoking method's own arguments, the additional instruction-level processing that sets up the invocation to the dispatch unit may also create an object made to contain the invoking method's own arguments (note that this same object may then be used by a plug-in module's handler to record the invoking method's arguments at the time the method starts). Regardless of how the arguments for the dispatch unit's methods are prepared, the procedures for doing so require the execution of instructions; and, therefore, modifying a bytecode-level method with invocations to another method typically requires additional instructions, per invocation, beyond each specific invocation instruction (e.g., beyond each added invokestatic instruction).
Because invocations to the dispatch unit are made from modified method entry and exit points,
Note that two different methods of the dispatch unit are invoked: dispatchunit.entry 643 a and dispatchunit.exit 644 a, 644 b. Here, the dispatchunit.entry method 643 a is invoked for those invocations made from a method entry point; and, the dispatchunit.exit method 644 a is invoked for those invocations made from a method exit point. Invoking different methods depending on whether the method is at an entry point or exit point allows for different or customized dispatch unit responses depending on whether the method resides at an entry or exit.
Other possible differences between the entry and exit method embodiments referred to in
3. object this pointer; and,
4. object [ ] param.
By contrast, the embodiments of the exit methods 644 a, 645 a include the following arguments:
3. returned value
4. thrown exception [which is normally null, and becomes non-null when the method has returned abruptly, i.e. by throwing an exception]
From the above note that both the entry and exit methods include an identification of the class that the invoking method 620 belongs to (classid) and an identification of the invoking method 620 itself (methodid). In an embodiment, the classid and methodid parameters are sufficient to identify the invoking method. In other embodiment, however, the classid, the methodid and a description of the invoking method's own arguments (which are referenced through an object identified by the “object [ ] param” argument) are used by the dispatch unit's entry method to identify the appropriate plug-in module treatment. As alluded to above with respect to
The .entry method and exit methods may differ, however, in that the .entry method may also identify (e.g., via the “object this pointer” argument) the object associated with the invoking method (noting that if the invoking method is a class (i.e. static) method the “object this pointer” argument may be represented as a null); yet, the exit method may make no such identification. As described above with respect to the embodiments of
The “object this pointer” parameter therefore does not need to be provided for .exit methods. The entry method and exit methods may also differ in that the entry method may describe the invoking method's own arguments as part of its invocation to the dispatch unit (i.e., in the case of the “GetMax” method, a and b as found in “object[param]”), whereas the .exit methods may only describe the returned output value as part of the invocation. This allows for different plug-in module treatment as between entry points and exit points of the same method. In one embodiment of the invention, if a method throws an exception, an exit method event may be processed.
Referring back to the bytecode instruction level coding 620 b, note that a third exit point and associated instruction level coding is also depicted 646 b. According to the approach alluded to in
In summary, the modification process 452 implemented by bytecode modifier 352, besides installing blocks of instructions 643 b, 644 b, 645 b for dispatching unit invocations at naturally existing method entry and exit points, may also be configured to install an artificial exit (i.e., one that would not exist were it not for the modification process) and associated instructions that are tailored to respond to an error condition, should one arise, during execution of the method. In such a case, the “i cmp” instruction (which is the instruction that signifies whether “a” is greater than “b” or if “b” is greater than “a”) triggers an error condition; which, in turn, would cause instructions 646 b to be executed.
Referring to both the bytecode level 620 b and source code level 620 a codings of the GetMax method, note that the error handling instructions 646 b correspond to an invocation to the dispatch unit 430. Here, an exit method is used to invoke the dispatch unit. Note that a “thrown” value (i.e., an error indication) is provided as part of the invocation. This allows a unique plug-in module to be utilized to in the case of an error condition. The error handling instructions 646 b also include a generic thrown command that, when executed, raises an error state in the current thread, which can either be caught or propagated up the execution stack.
The unmodified classfile 650 a in
The exemplary bytecode modified classfiile 650 b includes the following elements: (1) a “classid” field information structure 657; (2) a method information structure 656 for a method that registers with the dispatch unit when the modified classfile 650 b is loaded during runtime execution; (3) a method information structure 651 b for a modified version 653 of the constructor method; and (4) a method information structure 652 b for the modified version of the GetMax method 620 b that was discussed above with respect to
The additional classid field information structure 657 and the additional method information structure 656 (for the dispatch unit registration method) will be discussed together. Specifically, in an embodiment, when the modified classfile 650 b is loaded it automatically registers with the dispatch unit. In a further embodiment, the registration process entails providing the dispatch unit with the name of the class that the classfile 650 b represents, the name of its methods and the number and type of each method's arguments. In return, the dispatch unit provides the classfile with a numeric “classid” value that is to be used for invocations made to the dispatch unit during a runtime invocation. The classid value then becomes part of the classfile's constant pool. Accordingly, the additional method information structure 656 may include instructions for the registration method described just above; and, the classid information structure is used for the returned numeric “classid” parameter provided by the dispatch unit.
Viewing the constructor method prior to modification as a simple “straight through” method having a single entry point and a single exit point,
4.0. Dispatch Unit, Plug-In Modules and Handlers
Recall from the background that, in general, the runtime execution of any computer program can be viewed as the execution of a sequence of methods; and that, with respect to an object oriented approach, such a sequence of methods can be implemented by calling upon a sequence of objects and invoking one or more methods at each object. Referring back to
In one embodiment, the dispatch unit 330, 430 is implemented as a class whose entry and .exit methods are static methods of the dispatch unit class. In this same embodiment, each of the plug-in modules 331 1 through 331 N, 431 1 through 431 N are instantiated as separate objects from a “plug-in” class; and, the individual handlers of each plug-in module object correspond to its operable methods.
4.1. Dispatch Unit Implementation
The dispatch unit 430 may be implemented to refer to, as illustrated in
In response to the reception of the information, a method responsible for accessing the dictionary (e.g., a entry or a exit method) identifies the reference(s) that is listed for the invoking method; and, returns the reference(s) to the invoking method. Thus, from the perspective of an invoking method, an invocation is made to the dispatch unit 330, 430; which, in response, returns to the invoking method the reference to the appropriate plug-in module 331 1-331 N, 431 1-431 N. The process flow then jumps to the plug-in module(s), causing the corresponding handler instructions to be executed.
It is also worthwhile to point out that more than one plug-in module may be used for any single invocation from a modified method. For example, referring briefly back to
In this case, method 407 could still receive both tracing and coverage treatment by configuring the dictionary entry for method 407 to list both plug-in module A 460 (PIM_A) and plug-in module B 470 (PIM_B). By so doing, the process flow of
4.2. Extensions of Plugins and GUI Viewers
As mentioned above with respect to
The application trace plugin 810 of this embodiment identifies a particular application to the bytecode modifier 452 (e.g., in response to user input). In response, the bytecode modifier 452 modifies the entire application, by inserting function calls at entry points and exit points of each of the application's methods (e.g., using the bytecode modification techniques described herein).
Following modification, the application is executed, and the application trace handler 811 collects the timing data and, potentially, other method-related information, which it provides to a graphical user interface (“GUI”) 453. In one embodiment, tree generation logic 840 within the GUI 453 generates a method invocation tree 900 such as that illustrated in
The invocation tree 900 graphically displays information related to each invoked method, such as the time that it takes each method to complete and/or parameters used for each method, in a structured hierarchy. For the calculator application, three separate headings 901, 910, and 920 are generated, each of which identify a different thread of the calculator application (e.g., “SAPEngine_Application_Thread_5”). Entries for methods associated with each application thread are arranged beneath each respective thread heading. For example, method entry 902 is arranged directly beneath heading 901; method entries 911 are arranged directly beneath heading 910; and method entry 921 is arranged beneath heading 920.
Method entries 902, 911 and 921 represent “independent” methods in that they are not tied to the execution of another method (i.e., they are not invoked within the context of another method). By contrast, several “dependent” method entries 903-905, i.e., representing methods that are dependent on the method represented by entry 902, are graphically arranged under entry 902 within the invocation tree 900. For example, method entry 903 is dependent on method entry 902; method entry 904 is dependent on method entry 903; and method entries 905 are each dependent on method entry 904. As is known in the art, a first method is “dependent” on a second method if the first method is executed within the context of the second method. For example, the method represented by entry 903 is only executed within the method represented by entry 902. Thus, as indicated by dependency arrow 931, the dependency of methods increases as one moves towards the right within the invocation tree (i.e., methods towards the right are dependent on entries which are relatively further to the left).
In addition, as indicated by the time arrow 930 in
As mentioned above, various different types of information may be displayed for each entry within the invocation tree 900. In the example shown in
In one embodiment, a filtering module 841 is employed to filter timing data and other method-related information, based on the needs of the end user. For example, a user may choose to filter all method invocations from the method invocation tree 900 except for those methods associated with a particular application component, package or class. Similarly, in one embodiment, the filtering parameters may be applied so that only specific types of method-related information are displayed within the invocation tree 900. For example, the filtering parameters may specify that only each method's timing information is to be displayed whereas other information (e.g., the method's input/output variables, the counter value, . . . etc), is to be filtered out. Thus, the filtering module 841 provides a more precise level of control over the types of method's and the types of method-related information to be displayed within the graphical user interface 453.
An exemplary graphical user interface (“GUI”) for viewing information related to modified bytecode is illustrated in
Different information related to the bytecode modification may be displayed in the second window via a set of tabs 944. In
It should be noted that the specific GUI-based features described above are for the purpose of illustration only. Information related to bytecode modification may be displayed in a variety of additional and/or different ways while still complying with the underlying principles of the invention.
As described above, the application tracing plugin 810 instructs the bytecode modifier to modify all of an application's methods. While this may be sufficient for tracing/debugging a relatively small application, a higher level of precision may be desirable, particularly when working with large enterprise applications.
Thus, in contrast to the application tracing plugin 810 which causes the bytecode modifier 452 to modify all of the methods within a particular application, the user-configurable plugin 820 illustrated in
The method timing data and/or other method-related information may then be displayed within a method invocation tree 900 similar to that illustrated in
In addition, in one embodiment of the invention, a distributed statistical records (“DSR”) plugin 830 is employed to collect statistical data related to program execution across application servers, databases and/or external systems. Several specific examples of DSR tracing will now be described with respect to
In the specific example shown in
The client's 1000's initial request triggers a first method invocation 1001 within the application components 1011. The request may be, for example, a request for a particular record within the database. After additional application-layer processing, the application components 1011 generate a database request via method invocation 1003 which is received and processed by the database components 1021. A single request/response transaction with a client may involve multiple method-based interactions between the application server 1010 and the database server 1020. Thus, another method 1004 is invoked in response to a communication from the database server 1020; a response to the database server is generated as method invocation 1005 and a final response from the database server 1020 is received as method invocation 1006. Finally, the application responds to the client via method invocation 1002 (e.g., which may provide a Web page containing the requested data to the client).
As indicated in
The method-related information may also be provided to the GUI 453 which then generates an appropriate method invocation tree, and/or displays the method-related information in one or more additional ways (as described above). Although not illustrated in
The external system 1130 then processes the request and provides the results back to the application components via method invocation 1104. Finally, the application components respond to the client 1100 via method invocation 1002.
As in the examples shown in
The specific functions performed by each of the modules/components illustrated in
For the purpose of the present application,
Some end users may want to export the results of the bytecode modification techniques described herein to a standardized format, so that the results may be viewed and/or analyzed on other systems. To this end, as illustrated in
Instead of converting the invocation tree generated by the tree generation logic 840 into a markup language format, one embodiment of the markup language conversion module 1300 receives the method-related data directly from the a handler 811, 821, 831. This embodiment may be more efficient in that it circumvents the intermediate tree generation by the invocation tree generation logic 840.
5.0. Implementing Bytecode Modification
The discussions presented above focus upon characteristics of modified bytecode and the runtime execution flow of modified bytecode. The present section, by contrast, focuses on techniques for actually modifying the bytecode. Better said, referring back to
5.1 Processes For Modifying Classfiles
Referring to the exemplary classfile 103 b of
Likewise, the method information structure class 1502 can be used as a template to form separate “UPDATE ACCOUNTING”, “UPDATE BILLING” and “GetMax” objects 1502 1, 1502 2, 1502 3 that respectively represent the “UPDATE ACCOUNTING”, “UPDATE BILLING” and “GetMax” method information structures 111, 112, 131 found in the “sales” classfile 103 b.
Referring back to
It should be recognized however that if blocks of multiple instructions are to be inserted for each dispatch unit invocation (e.g., as suggested by instruction blocks 643 b, 6444 b, 645 b and 646 b of
According to the methodology of
The example referred to in
Specifically, an entire method information structure may be inserted into the classfile (such as a method information structure 656 for a dispatch unit registration method); and, a new field information structure may be inserted into the classfile (such as a field information structure 657 for a classid value). In keeping with the approach outlined above with respect to
Thus, whereas the modifications to the class methods themselves might involve modifying the collection of objects associated with each object made to represent a method, constructing a complete collection of objects that are representative of a completely modified classfile might involve the creation of objects having no counterpart to the pre-modified bytecode (e.g., the pre-modified bytecode has no “dispatch unit registration” method information structure nor “classid” field information structure).
Referring back to
5.2. Dispatch Unit Configuration
In one embodiment of the invention, in order for the dispatch unit 330, 430 to properly dispatch an invocation from a specific method to its appropriate plug-in during runtime, the dispatch unit is configured beforehand with the appropriate information. For example, before supporting an actual runtime flow of modified bytecode, the dispatch unit 430 may run through internal processes that “set up” information (e.g., a dictionary) that correlates each bytecode modified method with one or more appropriate plug-in modules. Recall that an embodiment of such a dictionary that correlated each modified bytecode method with an appropriate plug-in module has already been discussed with respect to
By way of example, as illustrated in
In an embodiment, registration occurs as a consequence of a modified classfile being “loaded.” The loading of a classfile can be viewed as the preparing of a class for subsequent use. In a typical case, a classfile is loaded before the classfile is used as a template for creating an object that is a member of the class that the classfile defines. In a further embodiment, a registration process is executed for a classfile before any of the classfile's modified methods are executed. So doing causes the dispatch unit 1702 to be properly updated for any subsequent invocation that arises from the execution of any of the class's modified methods.
The dispatch unit 1702 uses the class name, method names and method specific argument types received from any particular classfile as a stimulus for updating its internal information that correlates specific methods to specific plug-in modules (e.g., an internal dictionary). In various embodiments, the dispatch unit 1702 correlates class methods to their appropriate plug-in strictly numerically (i.e., without the use of character strings). For example, the dictionary of
In order to reference classes and methods strictly numerically, therefore, each character string based class name (e.g., “Sales”) is converted into a number (e.g., integer 1); and, each character string based method name (e.g., “GetMax”, “UPDATE ACCOUNTING”, “UPDATE BILLING”) is converted into a corresponding number (e.g., integer 1, integer 2, integer 3).
The numeric classids are configured into the dispatch unit's plug-in correlation dictionary; and, subsequent invocations made to the dispatch unit 1702 by way of a modified method are made to use the numeric classsid value rather than the given character string for the name. In further embodiments the numeric references are embodied as integers in order to improve processing speeds. Moreover, the numeric classid may be simply derived from the position at which a registering classfile registers with respect to a consecutive sequence of classfiles that register with the dispatch unit (e.g., the first classfile to register with the dispatch unit is given a classid of “1”, the second classfile to register with the dispatch unit is given a classid of “2”, etc.).
With respect to the conversion of character string method names into an appropriate number, in an embodiment, both the registering classsfiles 1701 1 through 1701 N and the dispatch unit 1702 are made to understand that methods are referred to numerically in a sequential order. For example, according to one approach, the methods are referred to by the order in which their corresponding information structures are listed in the classfile.
By making the classfiles and the dispatch unit understand the same ordered numeric naming scheme, methods can be properly referenced between the classfiles and the dispatch unit in a manner that is consistent with the method/plug-in-handler correlation entries that are maintained by the dispatch unit. For example, if classfile 1701 1 corresponds to the “Sales” classfile of
In an embodiment, during the registration process, a registering classfile sends method names in character string form to the dispatch unit 1702 in an order that allows the dispatch unit 1702 to properly deduce the correct numeric methodid values. For example, continuing with the above example, the “Sales” classfile 1701 1 would send method character string names in the following order: “GetMax”, “UPDATE ACCOUNTING”, “UPDATE BILLING”. By sending the character string names in the order listed above, both the registering classfile and the dispatch unit could independently recognize that the “GetMax” method is be given a methodid value of 1; the “UPDATE ACCOUNTING” method is to be given a value of 2; and, the “UPDATE BILLIG” method is to be given a value of 3.
Even though the dispatch unit 1702 can deduce the proper numeric naming scheme simply from knowing the order of the methods, explicit character strings are sent to the dispatch unit 1702 to also assist the dispatch unit in the updating of its method/plug-in module correlation scheme.
For example, referring back to
1) PlugIn A (460):
2) PlugIn B (470):
In an embodiment, the plug-in pattern 1811 is a body of information maintained (e.g., by way of an object) by the same class that the dispatch unit 1802 is associated with.
Once one or more appropriate plug-in modules have been identified by the dispatch unit's registration methodology 1807, the scheme used to correlate specific methods to their appropriate plug-in module(s) during runtime (e.g., dictionary 1812) is updated 1809 to account for all the methods of the registering classfile. The argument types for each method that are provided by a registering classfile can be added to the classid and methodid in each listing. If strictly numeric references are to be used for the classids and methodids, the character string-to-numeric conversion should be performed prior to the actual update 1809. For example, numeric methodids should be assigned based on the order of reception of method names from the registering classfile 1803; and, a suitable numeric classid should be identified. In an embodiment, the numeric classid is determined simply by incrementing a counter for each new registering classfile. The numeric classsid is then sent 1810 to the registering classfile 1801.
Given that the embodiment of the plug-in pattern described above identifies methods through their corresponding character strings, the dispatch unit's registration method 1807 would need a registering classfile 1801 to identify its methods with these same character strings so that the correct correlation between a modified method and its proper plug-in treatment could be carried out. Here, the use of character strings is emphasized because some form of “user” 1813 (e.g., a test or development engineer or maintenance Information Science (IS) personal) ultimately determines which methods are to receive which treatment from which plug-in modules.
For example, in the case of Distributed Statistical Records (DSR), typically, only a subset of eligible methods are selected for modification (e.g., “the important ones”). The intelligent selection of specific methods for modification ultimately boils down to their being listed by the plug-in pattern 1811. Accordingly, since methods and classes are most often dealt with by a user at a source code level or higher (e.g., GUI) with character strings—it is natural to expect that a “list” of methods to be treated as determined by a user would be listed in a character string format. Of course, in alternate embodiments, an initial list of character string names could be pre-formulated prior to registration so as to be represented purely in numeric form. This, in turn, may obviate registering with character strings.
Bytecode modification may be accomplished using a variety of different techniques and interfaces. As indicated in
As illustrated in
In addition, in one embodiment, a wizard-based GUI is provided for simplifying the bytecode modification process. Referring to
Embodiments of the invention may include various steps as set forth above. The steps may be embodied in machine-executable instructions which cause a general-purpose or special-purpose processor to perform certain steps. Alternatively, these steps may be performed by specific hardware components that contain hardwired logic for performing the steps, or by any combination of programmed computer components and custom hardware components.
Elements of the present invention may also be provided as a machine-readable medium for storing the machine-executable instructions. The machine-readable medium may include, but is not limited to, hard-disk drives, flash memory, optical disks, CD-ROMs, DVD ROMs, RAMs, EPROMs, EEPROMs, magnetic or optical cards, propagation media or other type of machine-readable media suitable for storing electronic instructions. For example, the present invention may be downloaded as a computer program which may be transferred from a remote computer (e.g., a server) to a requesting computer (e.g., a client) by way of data signals embodied in a carrier wave or other propagation medium via a communication link (e.g., a modem or network connection).
Throughout the foregoing description, for the purposes of explanation, numerous specific details were set forth in order to provide a thorough understanding of the invention. It will be apparent, however, to one skilled in the art that the invention may be practiced without some of these specific details. Accordingly, the scope and spirit of the invention should be judged in terms of the claims which follow.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5748963||May 12, 1995||May 5, 1998||Design Intelligence, Inc.||Adaptive binding|
|US6026237||Nov 3, 1997||Feb 15, 2000||International Business Machines Corporation||System and method for dynamic modification of class files|
|US6118940||Nov 25, 1997||Sep 12, 2000||International Business Machines Corp.||Method and apparatus for benchmarking byte code sequences|
|US6260187||Aug 20, 1998||Jul 10, 2001||Wily Technology, Inc.||System for modifying object oriented code|
|US6381735||Nov 20, 1998||Apr 30, 2002||Microsoft Corporation||Dynamic classification of sections of software|
|US6631515||Sep 24, 1998||Oct 7, 2003||International Business Machines Corporation||Method and apparatus to reduce code size and runtime in a Java environment|
|US6658600||Apr 24, 2000||Dec 2, 2003||Microsoft Corporation||Target control abstraction for debugging embedded systems|
|US6662359 *||Jul 20, 2000||Dec 9, 2003||International Business Machines Corporation||System and method for injecting hooks into Java classes to handle exception and finalization processing|
|US6895578||Jan 6, 2000||May 17, 2005||Parasoft Corporation||Modularizing a computer program for testing and debugging|
|US6934942 *||Aug 24, 2001||Aug 23, 2005||Microsoft Corporation||System and method for using data address sequences of a program in a software development tool|
|US6961918 *||Jun 21, 2001||Nov 1, 2005||Garner Robert E||System for intellectual property reuse in integrated circuit design|
|US6968540||Jun 12, 2001||Nov 22, 2005||Opnet Technologies Inc.||Software instrumentation method and apparatus|
|US7093234||Aug 24, 2001||Aug 15, 2006||International Business Machines Corporation||Dynamic CPU usage profiling and function call tracing|
|US20020073063 *||Aug 9, 2001||Jun 13, 2002||International Business Machines Corporation||Generation of runtime execution traces of applications and associated problem determination|
|US20020170036||Feb 21, 2002||Nov 14, 2002||Cobb Jeffrey R.||Detecting a stalled routine|
|US20040031020||Dec 31, 2002||Feb 12, 2004||International Business Machines Corporation||Method and system for modifying a class file to monitor data flow|
|US20040123279||Dec 18, 2002||Jun 24, 2004||International Business Machines Corporation||Method and system for auto-instrumenting java applications through probe injection|
|US20050039171||Aug 12, 2003||Feb 17, 2005||Avakian Arra E.||Using interceptors and out-of-band data to monitor the performance of Java 2 enterprise edition (J2EE) applications|
|US20050039187||Aug 12, 2003||Feb 17, 2005||Avakian Arra E.||Instrumenting java code by modifying bytecodes|
|1||Ajay Chander et al., "Mobile Code Security by Java Bytecode Instrumentation", Proceedings of the DARPA Information Survivability Conference & Exposition DISCEX-II 2001, Jun. 12-14, 2001, Stanford University and University of Pennsylvania, [*Partially supported by DARPA contract N66001-00-C-8015 and ONR grant N00014-97-1-0505] (14 pgs.).|
|2||Alan Snyder, "The Essence of Objects: Concepts and Terms", Jan. 1993, pp. 31-42, Sunsoft, Mountain View.|
|3||Alexander G. Shvets, "CafeBabe", 1999.|
|4||Alexander Keller, et al., "Measuring Application Response Times with the CIM Metrics Model", IBM e-Business, DSOM 2002, Session 2: Measuring Quality of Service, Montreal, Canada, Oct. 22, 2002.|
|5||Algis Rudys, et al., "Enforcing Java Run-Time Properties Using Bytecode Rewriting", International Symposium on Software Security (Tokyo, Japan), Nov. 2002, Rice University, Houston, TX 77005, USA (16 pgs.).|
|6||Allen Goldberg, et al., "Instrumentation of Java Bytecode for Runtime Analysis", Fifth ECOOP Workshop on Formal Techniques for Java-like Programs, Jul. 21, 2003, Kestrel Technology, NASA Ames Research Center, Moffett Field, California USA, (9 pgs.).|
|7||Alphaworks, "Jikes Bytecode Toolkit: Overview", www.alphaworks.ibm.com/tech/jikesbt, posted Mar. 31, 2000, 2 pages, printed Sep. 28, 2006.|
|8||Ben Stephenson, et al., "Characterization and Optimization of Java Applications" Department of Computer Science, Abstract in Western Research Forum Program & Abstracts, p. 20, 2003.|
|9||Duke University, "The Java Object Instrumentation Environment", www.cs.duke.edu/ari/joie/, last updated May 2003, printed Sep. 28, 2006, 2 pages.|
|10||Dylan McNamee et al., "Specialization Tools and Techniques for Systematic Optimization of System Software", Oregon Graduate Institute of Science & Technology, and University of Rennes/IRISA, ACM Transactions on Computer Systems, 2001 (30 pgs.).|
|11||Etienne Gagnon, et al., "Effective Inline-Threaded Interpretation of Java Bytecode Using Preparation Sequences", Sable Research Group, Université du Québec à Montréal and McGill University, Montreal. Canada, Jan. 2003 (15 pgs.).|
|12||Geoff A. Cohen et al., "Automatic Program Transformation with JOIE", Department of Computer Science, Duke University, 12 pages.|
|13||Geoff A. Cohen, et al., Software-Practice and Experience, [Version: Mar. 6, 2000 v2.1] "An Architecture for Safe Bytecode Insertion", Department of Computer Science, Duke University (27 pgs.).|
|14||Han Bok Lee, "BIT: Bytecode Instrumenting Tool" University of Colorado, Department of Computer Science 1997, 51 pages.|
|15||Han Bok Lee, et al., "BIT: A Tool for Instrumenting Java Bytecodes", originally published in the Proceedings of the USENIX Symposium on Internet Technologies and Systems, Monterey, California, Dec. 1997, www.usenix.org/ (11 pgs.).|
|16||Ian Welch, et al., "Kava-A Reflective Java Based on Bytecode Rewriting" SpringerLink -Verlag Berling Heidelberg 2000, Chapter, Lecture Notes in Computer Science, W. Cazzola, et al. Editors, Reflection and Software Engineering, LNCS, pp. 155-167.|
|17||Ingo Kegel, "JCLASSLIB Bytecode Viewer Help", EJ-Technologies GmBH, Aug. 20, 2003.|
|18||Jonathan Davies, et al., Proceedings of the 2nd international conference on "An Aspect Oriented Performance Analysis Environment", 10 pgs., 2003, Boston, Massachusetts Mar. 17-21, 2003.|
|19||Markus Dahm, "Welcome to the Byte Code Engineering Library 4.4.1", http://bcel.sourceforge.net/main.html, last updated Apr. 12, 2002, 2 pages, printed Sep. 28, 2006.|
|20||Mobile-Code Security Mechanisms for Jini-"Mobile-Code Security Mechanisms for Jini" Download code, DISCEX 2001 Paper, http://theory.stanford.edu/people/jcm/software/jinifilter.html, printed Jul. 2, 2004-(3 pgs.).|
|21||Nathan Macrides, Security Techniques for Mobile Code "SANS Security Essentials (GSEC) Practical Assignment Version 1.4", Jul. 11, 2002, (11 pgs.).|
|22||Nikolai G. Nikolov, "Execution of Modified Byte Code For Debugging, Testing And/Or Monitoring Of Object Oriented Software", U.S. Appl. No. 10/749,617, filed Dec. 30, 2003, Office Action mailed Jul. 10, 2007.|
|23||Nikolai G. Nikolov, "Registration Method For Supporting Bytecode Modification", U.S. Appl. No. 10/749,686, filed Dec. 30, 2003, Office Action mailed Jul. 3, 2007.|
|24||Peter W. Gill, "Probing for a Continued Validation Prototype", a Thesis Submitted to the Faculty of the Worcester Polytechnic Institute, May 2001, (111 pages).|
|25||Ralph Keller et al., "Supporting the Integration and Evolution of Components Through Binary Component Adaptation", www.cs.ucsb.edu/oocsb, Sep. 9, 1997, Technical Report TRCS97-15, 12 pages.|
|26||Reinhold Plösh, Johannes Kepler University Linz, Austria, "Evaluation of Assertion Support for the Java Programming Language", JOT: Journal of Object Technology, vol. 1, No. 3, Special issue: Tools USA 2002 Proceedings, pp. 5-17, http://www.jot.fm/issues/issue-2002-08/article1.|
|27||Reynald Affeldt, et al., "Supporting Objects in Run-Time Bytecode Specialization", Department of Graphics and Computer Science, University of Tokyo, ASIA-PEPM '02, Sep. 12-17, 2002, ACM, pp. 50-60.|
|28||Sun Microsystems, Java-J2EE 1.4 Application Server Developer's Guide, "Debugging J2EE Applications" Chapter 4, http://java.sun.com/j2ee/1.4/docs/devguide/dgdebug.html, 2003, printed Jul. 2, 2004 (11 pgs.).|
|29||Wen Li, et al., "Collaboration Transparency in the DISCIPLE Framework", CAIP Center, Rutgers-The State University of New Jersey, Piscataway, NJ, USA, Proceedings of the ACM International Conference on Supporting Group Work (Group '99) Nov. 14-17, 1999, Phoenix, AZ, (10 pgs).|
|30||Wily Technology, Inc., Wily Solutions "How Introscope(R) Works"-Enterprise Application Management, http/www.wilytech.com/solutions/products/howWorks.html, 1999-2004, printed Jul. 2, 2004 (1 page).|
|31||Wily Technology, Inc., Wily Solutions "Wily Introscope(R)"-Enterprise Application Management, http://www.wilytech.com/solutions/products/Introscope.html, 1999-2004, printed Jul. 2, 2004 (2 pgs.).|
|32||Wily Technology, Inc., Wily Technology, Inc., Wily Solutions "The Wily 5 Solution-Enterprise Applications are Your Business", http://www.wilytech.com/solutions/ibm-family.html, 1999-2004, printed Jul. 2, 2004 (2 pgs.).|
|33||www.cafeonleche.org/books/xmljava/chapters/ch09s04.html Chapter 9, The document Object Model, Trees, 2001, 2002, pp. 1-10.|
|34||www.cafeonleche.org/books/xmljava/chapters/ch09s05.html Chapter 9, The Document Object Model DOM Parsers for Java, 2001, 2002, pp. 1-3.|
|35||www.cafeonleche.org/books/xmljava/chapters/ch09x06.html Chapter 9, The Document Object Model Parsing documents with a DOM Parser, 2001, 2002, pp. 1-10.|
|36||www.cafeonleche.orge/books/xmljava/chapters/ch09.html Chapter 9, The Document Object Model, 2001, 2002, pp. 1-2.|
|37||www.caseonleche.org/books/xmljava/chapters/ch09s09.html Chapter 9, The Document Object Model JAXP Serialization, 2001, 2002, pp. 1-3.|
|38||www.w3.org/TR/DOM-Level-2-Core/introduction.html What is Document Object Model?, Nov. 13, 2000, pp. 1-7.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7707547 *||Mar 10, 2006||Apr 27, 2010||Aptana, Inc.||System and method for creating target byte code|
|US7817293 *||Oct 19, 2010||Infoprint Solutions Company, Llc||Trace and debug tool for high speed printer systems|
|US7950000 *||May 24, 2011||Microsoft Corporation||Architecture that restricts permissions granted to a build process|
|US8001521 *||Jun 12, 2006||Aug 16, 2011||Sap Ag||Meta-date driven implementation of business objects and their transactional behavior|
|US8117597 *||May 13, 2006||Feb 14, 2012||Shia So-Ming Daniel||Method and system for specifying and developing application systems with dynamic behavior|
|US8166077 *||Apr 24, 2012||International Business Machines Corporation||Mapping a class, method, package, and/or pattern to a component|
|US8201151 *||Jun 12, 2012||International Business Machines Corporation||Method and system for providing post-mortem service level debugging|
|US8261243 *||Aug 12, 2008||Sep 4, 2012||International Business Machines Corporation||Selective execution of trace mechanisms for applications having different bit structures|
|US8276117||Sep 25, 2012||International Business Machines Corporation||Displaying and refactoring programs that include database statements|
|US8285813||Dec 3, 2008||Oct 9, 2012||Appcelerator, Inc.||System and method for emulating different user agents on a server|
|US8291079||Jun 3, 2009||Oct 16, 2012||Appcelerator, Inc.||System and method for developing, deploying, managing and monitoring a web application in a single environment|
|US8402443 *||Aug 11, 2006||Mar 19, 2013||dyna Trace software GmbH||Method and system for automated analysis of the performance of remote method invocations in multi-tier applications using bytecode instrumentation|
|US8464224 *||Jun 30, 2009||Jun 11, 2013||Sap Ag||Integrated performance and load testing tool for application servers|
|US8473915 *||Aug 3, 2007||Jun 25, 2013||International Business Machines Corporation||Coverage analysis tool for testing database-aware software applications|
|US8527860||Dec 2, 2008||Sep 3, 2013||Appcelerator, Inc.||System and method for exposing the dynamic web server-side|
|US8527953 *||Jun 10, 2010||Sep 3, 2013||Accenture Global Services Limited||Assessment system for choosing maintenance approaches for GUI-directed test scripts|
|US8595697 *||Jul 1, 2009||Nov 26, 2013||International Business Machines Corporation||Serializing a templated markup language representation of test artifacts|
|US8719451||Nov 22, 2008||May 6, 2014||Appcelerator, Inc.||System and method for on-the-fly, post-processing document object model manipulation|
|US8756579||Nov 30, 2008||Jun 17, 2014||Appcelerator, Inc.||Client-side and server-side unified validation|
|US8806431||Dec 2, 2008||Aug 12, 2014||Appecelerator, Inc.||Aspect oriented programming|
|US8819539||Nov 30, 2008||Aug 26, 2014||Appcelerator, Inc.||On-the-fly rewriting of uniform resource locators in a web-page|
|US8880678||Jun 4, 2009||Nov 4, 2014||Appcelerator, Inc.||System and method for managing and monitoring a web application using multiple cloud providers|
|US8914774||Nov 14, 2008||Dec 16, 2014||Appcelerator, Inc.||System and method for tagging code to determine where the code runs|
|US8938491||Dec 2, 2008||Jan 20, 2015||Appcelerator, Inc.||System and method for secure binding of client calls and server functions|
|US8954553||Sep 20, 2009||Feb 10, 2015||Appcelerator, Inc.||System and method for developing, deploying, managing and monitoring a web application in a single environment|
|US9009672||Jul 25, 2013||Apr 14, 2015||Accenture Global Services Limited||Assessment system for choosing maintenance approaches for GUI-directed test scripts|
|US9141515 *||Jul 15, 2014||Sep 22, 2015||Sap Se||Limiting display content in editor for large data volumes|
|US9148467||Oct 5, 2012||Sep 29, 2015||Appcelerator, Inc.||System and method for emulating different user agents on a server|
|US9378034 *||Dec 20, 2013||Jun 28, 2016||Sap Se||Dynamic bytecode modification of classes and class hierarchies|
|US20050210448 *||Mar 17, 2004||Sep 22, 2005||Kipman Alex A||Architecture that restricts permissions granted to a build process|
|US20060152754 *||Jan 7, 2005||Jul 13, 2006||Boyd Erin A||Trace and debug tool for high speed printer systems|
|US20060230070 *||Mar 10, 2006||Oct 12, 2006||Xamlon, Inc.||System and method for creating target byte code|
|US20060259289 *||May 13, 2006||Nov 16, 2006||Shia So-Ming D||Method and system for specifying and developing application systems with dynamic behavior|
|US20070169055 *||Aug 11, 2006||Jul 19, 2007||Bernd Greifeneder||Method and system for automated analysis of the performance of remote method invocations in multi-tier applications using bytecode instrumentation|
|US20070288891 *||Jun 12, 2006||Dec 13, 2007||Michael Aakolk||Meta-data driven implementation of business objects and their transactional behavior|
|US20090037873 *||Aug 3, 2007||Feb 5, 2009||Azadeh Ahadian||Displaying and refactoring programs that include database statements|
|US20090037893 *||Aug 3, 2007||Feb 5, 2009||Stephen Andrew Brodsky||Coverage analysis tool for database-aware applications|
|US20090164978 *||Dec 20, 2007||Jun 25, 2009||International Business Machines Corporation||Method and system for providing post-mortem service level debugging|
|US20100042977 *||Feb 18, 2010||International Business Machines Corporation||Selective Execution of Trace Mechanisms for Applications Having Different Bit Structures|
|US20100083228 *||Sep 30, 2008||Apr 1, 2010||International Business Machines Corporation||Mapping a Class, Method, Package, and/or Pattern to a Component|
|US20100318970 *||Dec 16, 2010||Mark Grechanik||Assessment system for choosing maintenance approaches for gui-directed test scripts|
|US20100333072 *||Jun 30, 2009||Dec 30, 2010||Sap Ag||Integrated performance and load testing tool for application servers|
|US20110004865 *||Jan 6, 2011||International Business Machines Corporation||Serializing a templated markup language representation of test artifacts|
|US20140344792 *||Dec 20, 2013||Nov 20, 2014||Sap Ag||Dynamic bytecode modification of classes and class hierarchies|
|U.S. Classification||717/128, 717/130, 717/131, 717/148|
|International Classification||G06F9/45, G06F9/44|
|Cooperative Classification||G06F11/3612, G06F9/44526|
|European Classification||G06F11/36A4, G06F9/445L2|
|Jul 6, 2004||AS||Assignment|
Owner name: SAP AKTIENGESELLSCHAFT, GERMANY
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:KABADIYSKI, MARIO;NIKOLOV, NIKOLAI G.;REEL/FRAME:014823/0075
Effective date: 20040702
|Aug 24, 2012||FPAY||Fee payment|
Year of fee payment: 4
|Aug 26, 2014||AS||Assignment|
Free format text: CHANGE OF NAME;ASSIGNOR:SAP AG;REEL/FRAME:033625/0334
Owner name: SAP SE, GERMANY
Effective date: 20140707